{"id":169,"date":"2017-01-16T20:09:04","date_gmt":"2017-01-16T12:09:04","guid":{"rendered":"http:\/\/www.mrtblog.cn\/?p=169"},"modified":"2023-03-05T14:09:17","modified_gmt":"2023-03-05T06:09:17","slug":"%e5%89%8d%e8%a8%80","status":"publish","type":"post","link":"http:\/\/www.mrtblog.cn\/?p=169","title":{"rendered":"sklearn\u5b66\u4e60"},"content":{"rendered":"
<\/span> 1,231<\/span> Views<\/span><\/div>

1.\u524d\u8a00<\/h1>\n

\u6b63\u503c\u5bd2\u5047\u65f6\u671f\uff0c\u501f\u6b64\u673a\u4f1a\u5b66\u4e60sklearn\uff0c\u540c\u65f6\u8bb0\u5f55\u81ea\u5df1\u7684\u5b66\u4e60\u7ecf\u5386\uff0c\u4ee5\u4fbf\u5c06\u6765\u590d\u4e60\u3002<\/h4>\n

Scikit-learn\u5b66\u4e60\u5fc5\u8981\u6761\u4ef6\uff1a
\nPython\uff08> = 2.6\u6216> = 3.3\uff09\uff0c
\nNumPy\uff08> = 1.6.1\uff09\uff0c
\nSciPy\uff08> = 0.9\uff09\u3002
\n\u53c2\u8003windows\u4e0bpython\u5b89\u88c5NumPy\u548cSciPy\u6a21\u5757<\/a><\/p>\n

\u4e0b\u8f7d\u5730\u5740\uff1a
\nPython\uff1ahttps:\/\/www.python.org\/<\/a>
\nNumPy\uff1ahttp:\/\/www.numpy.org\/<\/a>
\nSciPy\uff1ahttp:\/\/www.scipy.org\/<\/a>
\nScikit-learn\uff1ahttp:\/\/scikit-learn.org\/stable\/<\/a>
\nMatploylib\uff1ahttp:\/\/matplotlib.org\/<\/a>
\n\u6839\u636e\u81ea\u8eab\u5b9e\u9645\u60c5\u51b5\uff0c\u6240\u6709\u6587\u4ef6\u90fd\u80fd\u5728\u8fd9\u91cc\u4e0b\u8f7d\uff1a
\nhttp:\/\/www.lfd.uci.edu\/~gohlke\/pythonlibs\/<\/a><\/p>\n

\u7cfb\u7edf\u73af\u5883\uff1a
\nWindows\uff1a64\u4f4d 4\u6838 6GB RAM
\nPython\uff1a64\u4f4d 2.7.13
\nNumPy\uff1anumpy-1.11.3+mkl-cp27-cp27m-win_amd64\uff08\u6ce8\u610f\u542b\u6709mkl\uff0c\u5426\u5219\u5bfc\u5165sklearn\u90e8\u5206\u5305\u4f1a\u51fa\u9519\uff09
\nSciPy\uff1ascipy-0.19.0-cp27-cp27m-win_amd64
\nScikit-learn\uff1ascikit_learn-0.18.1-cp27-cp27m-win_amd64
\nMatplotlib\uff1amatplotlib-1.5.3-cp27-cp27m-win_amd64
\n\u76f8\u5173\u6559\u7a0b\u7f51\u4e0a\u6709\u5f88\u591a\uff0c\u5927\u81f4\u770b\u4e86\u4e0bnumpy\uff0c\u5176\u4ed6\u5185\u5bb9\u7b49\u5b9e\u9645\u7528\u5230\u7684\u65f6\u5019\u518d\u9488\u5bf9\u5b66\u4e60\u3002<\/p>\n

2.\u4f7f\u7528scikit-learn\u8fdb\u884c\u673a\u5668\u5b66\u4e60\u7684\u4ecb\u7ecd<\/h1>\n

\u7ffb\u8bd1\u81eahttp:\/\/scikit-learn.org\/stable\/tutorial\/basic\/tutorial.html<\/a>\uff0c\u5185\u5bb9\u6709\u5220\u6539<\/p>\n

\u7ae0\u8282\u5185\u5bb9<\/h2>\n

\u5728\u672c\u8282\u4e2d\u4ecb\u7ecd\u5728scikit-learn\u4e2d\u4f7f\u7528\u7684\u673a\u5668\u5b66\u4e60\u8bcd\u6c47\uff0c\u5e76\u7ed9\u51fa\u4e00\u4e2a\u7b80\u5355\u7684\u5b66\u4e60\u793a\u4f8b\u3002<\/p>\n

\u673a\u5668\u5b66\u4e60\uff1a\u95ee\u9898\u8bbe\u7f6e<\/h2>\n

\u4e00\u822c\u6765\u8bf4\uff0c\u5b66\u4e60\u95ee\u9898\u8003\u8651\u4e00\u7ec4n\u4e2a\u6570\u636e\u6837\u672c\uff0c\u7136\u540e\u5c1d\u8bd5\u9884\u6d4b\u672a\u77e5\u6570\u636e\u7684\u5c5e\u6027\u3002\u5982\u679c\u6bcf\u4e2a\u6837\u672c\u4e0d\u4ec5\u4ec5\u662f\u5355\u4e2a\u6570\u5b57\uff0c\u5e76\u4e14\u6709\u4f8b\u5982\u591a\u7ef4\u6761\u76ee\uff08\u4e5f\u79f0\u4e3a\u591a\u53d8\u91cf\u6570\u636e\uff09\uff0c\u5219\u79f0\u5176\u5177\u6709\u82e5\u5e72\u5c5e\u6027\u6216\u7279\u5f81\u3002
\n\u6211\u4eec\u53ef\u4ee5\u628a\u5b66\u4e60\u95ee\u9898\u5212\u5206\u4e3a\u51e0\u4e2a\u5927\u7684\u79cd\u7c7b\uff1a<\/p>\n

\u76d1\u7763\u5b66\u4e60\uff1a<\/strong>\u6570\u636e\u5e26\u6709\u6211\u4eec\u60f3\u8981\u9884\u6d4b\u7684\u9644\u52a0\u5c5e\u6027\uff0c\u76d1\u7763\u5b66\u4e60\u5305\u542b\uff1a<\/p>\n

\u5206\u7c7b\uff1a<\/strong>\u6837\u672c\u5c5e\u4e8e\u4e24\u4e2a\u6216\u66f4\u591a\u4e2a\u7c7b\uff0c\u5e76\u4e14\u6211\u4eec\u60f3\u4ece\u5df2\u7ecf\u6807\u8bb0\u7684\u6570\u636e\u4e2d\u5b66\u4e60\u5982\u4f55\u9884\u6d4b\u672a\u6807\u8bb0\u6570\u636e\u7684\u7c7b\u3002\u5206\u7c7b\u95ee\u9898\u7684\u4e00\u4e2a\u4f8b\u5b50\u662f\u624b\u5199\u6570\u5b57\u7684\u8bc6\u522b\uff0c\u5176\u4e2d\u76ee\u7684\u662f\u5c06\u6bcf\u4e2a\u8f93\u5165\u5411\u91cf\u5206\u914d\u7ed9\u6570\u91cf\u6709\u9650\u7684\u79bb\u6563\u7c7b\u522b\u4e2d\u7684\u4e00\u4e2a\u3002\u53e6\u4e00\u79cd\u7406\u89e3\u5206\u7c7b\u7684\u65b9\u5f0f\u662f\u5c06\u5176\u4f5c\u4e3a\u76d1\u7763\u5b66\u4e60\u7684\u79bb\u6563\uff08\u800c\u4e0d\u662f\u8fde\u7eed\uff09\u5f62\u5f0f\uff0c\u4e00\u4e2a\u662f\u8ba9\u63d0\u4f9b\u7684\u6bcf\u4e00\u4e2an\u6837\u672c\u5177\u6709\u6709\u9650\u6570\u91cf\u7684\u7c7b\u522b\uff0c\u4e00\u4e2a\u662f\u5c1d\u8bd5\u7528\u6b63\u786e\u7684\u7c7b\u522b\u6216\u5206\u7ec4\u6765\u6807\u8bb0\u5b83\u4eec \u3002<\/p>\n

\u56de\u5f52\uff1a<\/strong>\u5982\u679c\u671f\u671b\u7684\u8f93\u51fa\u7531\u4e00\u4e2a\u6216\u591a\u4e2a\u8fde\u7eed\u53d8\u91cf\u7ec4\u6210\uff0c\u5219\u4efb\u52a1\u79f0\u4e3a\u56de\u5f52\u3002\u56de\u5f52\u95ee\u9898\u7684\u4e00\u4e2a\u4f8b\u5b50\u662f\u9c91\u9c7c\u7684\u957f\u5ea6\u4f5c\u4e3a\u51fd\u6570\u9884\u6d4b\u5e74\u9f84\u548c\u4f53\u91cd\u3002<\/p>\n

\u65e0\u76d1\u7763\u5b66\u4e60\uff1a<\/strong>\u8bad\u7ec3\u6570\u636e\u7531\u4e00\u7ec4\u6ca1\u6709\u4efb\u4f55\u76f8\u5e94\u76ee\u6807\u503c\u7684\u8f93\u5165\u5411\u91cfx\u7ec4\u6210\u3002\u95ee\u9898\u7684\u76ee\u6807\u53ef\u4ee5\u662f\u5c06\u6570\u636e\u4e2d\u53d1\u73b0\u76f8\u4f3c\u7684\u5206\u7ec4\uff0c\u79f0\u4e3a\u805a\u7c7b<\/strong>\uff1b\u6216\u8005\u786e\u5b9a\u5728\u8f93\u5165\u7a7a\u95f4\u5185\u7684\u6570\u636e\u7684\u5206\u5e03\uff0c\u79f0\u4e3a\u5bc6\u5ea6\u4f30\u8ba1<\/strong>\uff1b\u6216\u8005\u5c06\u6570\u636e\u4ece\u9ad8\u7ef4\u7a7a\u95f4\u4e0b\u964d\u52302\u7ef4\u62163\u7ef4\u4ee5\u4fbf\u4e8e\u53ef\u89c6\u5316\u3002<\/p>\n

\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6<\/h3>\n

\u673a\u5668\u5b66\u4e60\u662f\u5173\u4e8e\u5b66\u4e60\u5df2\u77e5\u6570\u636e\u96c6\u7684\u4e00\u4e9b\u5c5e\u6027\u5e76\u5c06\u5b83\u4eec\u5e94\u7528\u5230\u65b0\u6570\u636e\u3002\u8fd9\u5c31\u662f\u4e3a\u4ec0\u4e48\u5728\u673a\u5668\u5b66\u4e60\u8bc4\u4f30\u7b97\u6cd5\u7684\u5e38\u89c1\u505a\u6cd5\u662f\u5c06\u624b\u5934\u7684\u6570\u636e\u5206\u6210\u4e24\u7ec4\uff0c\u4e00\u7ec4\u6211\u4eec\u79f0\u4e3a\u8bad\u7ec3\u96c6\uff0c\u6211\u4eec\u5728\u5176\u4e0a\u5b66\u4e60\u6570\u636e\u5c5e\u6027\uff0c\u4e00\u7ec4\u6211\u4eec\u79f0\u4e3a\u6d4b\u8bd5\u96c6\uff0c\u6211\u4eec\u6d4b\u8bd5\u8fd9\u4e9b\u5c5e\u6027\u3002<\/p>\n

\u52a0\u8f7d\u793a\u4f8b\u6570\u636e\u96c6<\/h2>\n

scikit-learn\u9644\u5e26\u4e86\u4e00\u4e9b\u6807\u51c6\u6570\u636e\u96c6\uff0c\u4f8b\u5982\u7528\u4e8e\u5206\u7c7b\u7684iris<\/a>\u6570\u636e\u96c6\u548cdigit<\/a>\u6570\u636e\u96c6\u4ee5\u53ca\u7528\u4e8e\u56de\u5f52\u7684boston house prices<\/a>\u6570\u636e\u96c6\u3002<\/p>\n

\u63a5\u4e0b\u6765\uff0c\u4f7f\u7528Python\u89e3\u91ca\u5668\uff0c\u52a0\u8f7diris\u548cdigits\u6570\u636e\u96c6\uff1a<\/p>\n

\nfrom sklearn import datasets\niris=datasets.load_iris()\ndigits=datasets.load_digits()\n<\/code><\/pre>\n
\u6570\u636e\u96c6\u662f\u4e00\u4e2a\u7c7b\u4f3c\u5b57\u5178\u7684\u5bf9\u8c61\uff0c\u5b83\u4fdd\u5b58\u6240\u6709\u6570\u636e\u548c\u4e00\u4e9b\u6709\u5173\u6570\u636e\u7684\u5143\u6570\u636e\u3002\u6b64\u6570\u636e\u5b58\u50a8\u5728.data\u6210\u5458\u4e2d\uff0c\u8fd9\u662f\u4e00\u4e2an_sample\uff0cn_feature\uff08n\u4e2a\u6837\u672c\uff0cn\u4e2a\u7279\u5f81\uff09\u7684\u6570\u7ec4\u3002\u5728\u6709\u76d1\u7763\u7684\u60c5\u51b5\u4e0b\uff0c\u4e00\u4e2a\u6216\u591a\u4e2a\u76f8\u5e94\u53d8\u91cf\u5b58\u50a8\u5728.target\u6210\u5458\u4e2d\u3002\u4f8b\u5982\uff0c\u5728\u4f7f\u7528digit\u6570\u636e\u96c6\u7684\u60c5\u51b5\u4e0b\uff0cdigits.data\u5141\u8bb8\u8bbf\u95ee\u53ef\u7528\u4e8e\u5bf9\u6570\u5b57\u6837\u672c\u8fdb\u884c\u5206\u7c7b\u7684\u7279\u5f81\uff0cdigits.target\u7ed9\u51fa\u4e86\u6570\u5b57\u6570\u636e\u96c6\u7684\u771f\u5b9e\u6807\u6ce8\uff08\u8be5\u6570\u636e\u5c5e\u4e8e\u54ea\u4e2a\u7c7b\u522b\uff09\uff0c\u5373\u5bf9\u5e94\u4e8e\u6211\u4eec\u8bd5\u56fe\u5b66\u4e60\u7684\u6bcf\u4e2a\u6570\u5b57\u56fe\u50cf\u7684\u6570\u5b57\uff1a<\/p>\n
\ndigits.data\ndigits.target\n<\/code><\/pre>\n
\u4e0b\u56fe\u4e3a\u6d4b\u8bd5\u8f93\u51fadigits.data\u548cdigits.target<\/p>\n
\"\"<\/a><\/p>\n
\u6570\u7ec4\u7684shape\uff08\u5f62\u72b6\uff09\u5c5e\u6027<\/h3>\n
\u5c3d\u7ba1\u539f\u59cb\u6570\u636e\u53ef\u80fd\u5177\u6709\u4e0d\u540c\u7684\u5f62\u72b6\uff0c\u6570\u636e\u59cb\u7ec8\u662f\u4e8c\u7ef4\u6570\u7ec4\uff0c\u4e14\u5f62\u72b6\u4e3a\uff08n_samples\uff0cn_features\uff09\u3002\u8003\u8651digit\u6570\u636e\u96c6\uff0c\u6bcf\u4e2a\u539f\u59cb\u6837\u672c\u662fshape\uff088,8\uff09\u7684\u56fe\u50cf\uff0c\u5e76\u4e14\u53ef\u4ee5\u4f7f\u7528\u4ee5\u4e0b\u8bbf\u95ee\uff1a<\/p>\n
\ndigits.images[0]\n<\/code><\/pre>\n
<\/a>
\n\u8fd9\u4e2a\u6837\u4f8b\u5c55\u793a\u4e86\u5982\u4f55\u4ece\u539f\u59cb\u95ee\u9898\u5f00\u59cb\uff0c\u901a\u8fc7sklearn\u5851\u9020\u6570\u636e\u5e76\u4f7f\u7528\u3002<\/p>\n
\u5b66\u4e60\u53ca\u9884\u6d4b<\/h3>\n
\u4f7f\u7528digit\u6570\u636e\u96c6\uff0c\u7ed9\u5b9a\u56fe\u50cf\uff0c\u9884\u6d4b\u5176\u8868\u793a\u54ea\u4e2a\u6570\u5b57\u3002\u6211\u4eec\u7ed9\u51fa\u4e8610\u4e2a\u53ef\u80fd\u7684\u7c7b\uff08\u6570\u5b570\u52309\uff09\u4e2d\u7684\u6bcf\u4e00\u4e2a\u7684\u6837\u672c\uff0c\u8c03\u6574\u8bc4\u4f30\u5668\u4ee5\u80fd\u591f\u9884\u6d4b\u672a\u77e5\u6837\u672c\u6240\u5c5e\u7684\u7c7b\u3002
\n\u5728scikit-learn\u4e2d\uff0c\u5206\u7c7b\u7684\u8bc4\u4f30\u5668\u662f\u5b9e\u73b0\u65b9\u6cd5fit\uff08x\uff0cy\uff09\u548cpredict\uff08T\uff09\u7684Python\u5bf9\u8c61\u3002
\n\u8bc4\u4f30\u5668\u7684\u4e00\u4e2a\u793a\u4f8b\u662f\u5b9e\u73b0\u652f\u6301\u5411\u91cf\u673a\uff08SVM\uff09\u5206\u7c7b\u7684\u7c7bsklearn.svm.SVC\u3002\u8bc4\u4f30\u5668\u7684\u6784\u9020\u51fd\u6570\u91c7\u7528\u6a21\u578b\u7684\u53c2\u6570\u4f5c\u4e3a\u53c2\u6570\uff0c\u4f46\u6682\u65f6\uff0c\u6211\u4eec\u5c06\u628a\u8bc4\u4f30\u5668\u89c6\u4e3a\u4e00\u4e2a\u9ed1\u76d2\uff1a<\/p>\n
\nfrom sklearn import svm\nclf = svm.SVC(gamma=0.001, C=100.)\n<\/code><\/pre>\n
\u9009\u62e9\u6a21\u578b\u7684\u53c2\u6570<\/strong>
\n\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u624b\u52a8\u8bbe\u7f6egamma\u7684\u503c\u3002\u901a\u8fc7\u4f7f\u7528\u8bf8\u5982\u7f51\u683c\u641c\u7d22\u548c\u4ea4\u53c9\u9a8c\u8bc1\u7684\u5de5\u5177\uff0c\u53ef\u4ee5\u81ea\u52a8\u5730\u4e3a\u53c2\u6570\u627e\u5230\u597d\u7684\u503c\u3002
\n\u521b\u5efa\u7684\u8bc4\u4f30\u5668\u5b9e\u4f8bclf\uff0c\u5b83\u662f\u4e00\u4e2a\u5206\u7c7b\u5668\uff0c\u5fc5\u987b\u4ece\u6a21\u578b\u4e2d\u5b66\u4e60\uff0c\u8fd9\u662f\u901a\u8fc7\u5c06\u6211\u4eec\u7684\u8bad\u7ec3\u96c6\u4f20\u9012\u7ed9fit\u65b9\u6cd5\u6765\u5b8c\u6210\u3002\u4f5c\u4e3a\u8bad\u7ec3\u96c6\uff0c\u6211\u4eec\u4f7f\u7528\u9664\u4e86\u6700\u540e\u4e00\u4e2a\u6570\u636e\u96c6\u4e4b\u5916\u7684\u6240\u6709\u6570\u636e\u96c6\u7684\u56fe\u50cf\u3002\u7528[:-1] Python\u8bed\u6cd5\u9009\u62e9\u8fd9\u4e2a\u8bad\u7ec3\u96c6\uff0c\u5b83\u4ea7\u751f\u4e00\u4e2a\u65b0\u6570\u7ec4\uff0c\u5176\u4e2d\u5305\u542bdigits.data\u7684\u6700\u540e\u4e00\u4e2a\u6761\u76ee\uff1a<\/p>\n
\nclf.fit(digits.data[:-1], digits.target[:-1])\n<span style="color: #ff0000;">OUTPUT\uff1a<\/span>\nSVC(C=100.0, cache_size=200, class_weight=None, coef0=0.0,\ndecision_function_shape=None, degree=3, gamma=0.001, kernel='rbf',\nmax_iter=-1, probability=False, random_state=None, shrinking=True,\ntol=0.001, verbose=False)\n<\/code><\/pre>\n
\u73b0\u5728\u4f60\u53ef\u4ee5\u95ee\u5206\u7c7b\u5668\u5728digits\u6570\u636e\u96c6\u4e2d\u6700\u540e\u4e00\u4e2a\u56fe\u50cf\u7684\u6570\u5b57\u662f\u4ec0\u4e48\uff08\u6700\u540e\u8fd9\u4e00\u9879\u6ca1\u6709\u7528\u6765\u8bad\u7ec3\u5206\u7c7b\u5668\uff09\uff1a<\/p>\n
<\/a>
\n\u53ef\u4ee5\u770b\u5230\u5373\u4f7f\u56fe\u50cf\u5206\u8fa8\u7387\u5f88\u5dee\uff0c\u4f46\u662f\u5206\u7c7b\u8bc4\u4f30\u5668\u7ed9\u51fa\u7684\u7b54\u6848\u8fd8\u662f\u4ee4\u4eba\u4fe1\u670d\u7684\u3002
\n\u6a21\u578b\u6301\u4e45\u6027<\/strong>
\n\u5728\u4e4b\u540e\u76843.4\u90e8\u5206\u4ed4\u7ec6\u5b66\u4e60\uff0c\u7565<\/p>\n
\u9ed8\u8ba4\u89c4\u5219<\/h3>\n
scikit-learn\u8bc4\u4f30\u5668\u9075\u5faa\u67d0\u4e9b\u89c4\u5219\uff0c\u4f7f\u4ed6\u4eec\u7684\u884c\u4e3a\u66f4\u5177\u9884\u6d4b\u6027\u3002
\n\u9ed8\u8ba4\u6570\u636e\u7c7b\u578b<\/strong>
\n\u9664\u975e\u53e6\u6709\u8bf4\u660e\uff0c\u5426\u5219\u8f93\u5165\u5c06\u88ab\u8f6c\u6362\u4e3afloat64
\n\u4fee\u6539\u548c\u66f4\u65b0\u53c2\u6570<\/strong>
\n\u5728\u901a\u8fc7sklearn.pipeline.Pipeline.set_params\u65b9\u6cd5\u6784\u9020\u4e4b\u540e\uff0c\u53ef\u4ee5\u66f4\u65b0\u8bc4\u4f30\u5668\u7684\u53c2\u6570\u3002\u591a\u6b21\u8c03\u7528fit\uff08\uff09\u4f1a\u8986\u76d6\u4efb\u4f55\u4ee5\u524d\u7684fit\uff08\uff09\u7684\u5185\u5bb9\u3002
\n\u591a\u7c7b\u4e0e\u591a\u6807\u7b7e\u62df\u5408<\/strong>
\n\u5f53\u4f7f\u7528\u591a\u7c7b\u5206\u7c7b\u5668\u65f6\uff0c\u6267\u884c\u7684\u5b66\u4e60\u548c\u9884\u6d4b\u4efb\u52a1\u53d6\u51b3\u4e8e\u9002\u5408\u7684\u76ee\u6807\u6570\u636e\u7684\u683c\u5f0f<\/p>\n
3.\u9488\u5bf9\u79d1\u5b66\u6570\u636e\u5904\u7406\u7684\u7edf\u8ba1\u5b66\u4e60\u6559\u7a0b<\/h1>\n
\u53d1\u73b0\u7f51\u4e0a\u5df2\u7ecf\u6709\u4eba\u7ffb\u8bd1\u5b98\u65b9\u6587\u732e\u4e86\uff0c\u5728\u8fd9\u4e2a\u57fa\u7840\u4e0a\u7a0d\u4f5c\u4fee\u6539\uff0c\u8ba9\u7ffb\u8bd1\u53d8\u5f97\u9002\u5408\u81ea\u5df1\u7684\u98ce\u683c\u3002
\n\u5b98\u65b9\u539f\u6587\uff1a
\nhttp:\/\/scikit-learn.org\/stable\/tutorial\/statistical_inference\/index.html<\/a>
\n\u7ffb\u8bd1\u539f\u6587\uff1a
\nhttp:\/\/www.cnblogs.com\/taceywong\/p\/4570155.html<\/a>  \u4e2a\u4eba\u6839\u636e\u5b9e\u9645\u60c5\u51b5\u6709\u5220\u6539<\/p>\n
\u4e00\u3001\u7edf\u8ba1\u5b66\u4e60\uff1ascikit-learn\u4e2d\u7684\u8bbe\u7f6e\u4e0e\u8bc4\u4f30\u51fd\u6570\u5bf9\u8c61<\/h2>\n
\uff081\uff09\u6570\u636e\u96c6<\/h3>\n
scikit-learn \u4ece\u4e8c\u7ef4\u6570\u7ec4\u63cf\u8ff0\u7684\u6570\u636e\u4e2d\u5b66\u4e60\u4fe1\u606f\u3002\u4ed6\u4eec\u53ef\u4ee5\u88ab\u7406\u89e3\u6210\u591a\u7ef4\u89c2\u6d4b\u6570\u636e\u7684\u5217\u8868\u3002\u5982\uff08n,m\uff09\uff0cn\u8868\u793a\u6837\u4f8b\u8f74\uff0cy\u8868\u793a\u7279\u5f81\u8f74\u3002<\/p>\n
\u4f7f\u7528scikit-learn\u52a0\u8f7d\u4e00\u4e2a\u7b80\u5355\u7684\u6837\u4f8b\uff1airis\u6570\u636e\u96c6<\/strong>\uff08\u76f8\u540c\u7684\u5185\u5bb9\u4e0d\u622a\u56fe\u4e86\uff0c\u4e0a\u56de\u5df2\u7ecf\u622a\u8fc7\uff09<\/p>\n
\n>>>from sklearn import datasets\n>>>iris = datasets.load_iris()\n>>>data = iris.data \n>>>data.shape\n(150, 4)\n<\/code><\/pre>\n
\u5b83\u6709150\u4e2airis\u89c2\u6d4b\u6570\u636e\u6784\u6210\uff0c\u6bcf\u4e00\u4e2a\u6837\u4f8b\u6709\u56db\u4e2a\u7279\u5f81\uff1a\u843c\u7247\u3001\u82b1\u74e3\u957f\u5ea6\u3001\u82b1\u74e3\u5bbd\u5ea6\uff1b\u5177\u4f53\u7684\u4fe1\u606f\u53ef\u4ee5\u901a\u8fc7iris.DESCR\u67e5\u770b\u3002<\/p>\n
\u5f53\u6570\u636e\u521d\u59cb\u65f6\u4e0d\u662f(n\u6837\u4f8b\uff0cn\u7279\u5f81)\u6837\u5f0f\u65f6\uff0c\u9700\u8981\u5c06\u5176\u9884\u5904\u7406\u4ee5\u88abscikit-learn\u4f7f\u7528\u3002<\/p>\n
\u901a\u8fc7\u6570\u5b57\u6570\u636e\u96c6\u8bf4\u660e\u6570\u636e\u53d8\u5f62<\/strong>
\ndigits\u6570\u636e\u96c6\u75311797\u4e2a8×8\u624b\u5199\u6570\u5b57\u56fe\u7247\u7ec4\u6210<\/p>\n
\n>>>digits = datasets.load_digits()\n>>>digits.images.shape\n(1797, 8, 8) \n>>> import pylab as pl \n>>> pl.imshow(digits.images[-1], cmap=pl.cm.gray_r)\n<matplotlib.image.AxesImage object at ...>\n<\/code><\/pre>\n

\n\u5728scikit-learn\u4e2d\u4f7f\u7528\u8fd9\u4e2a\u6570\u636e\u96c6\uff0c\u6211\u4eec\u9700\u8981\u5c06\u5176\u6bcf\u4e00\u4e2a8×8\u56fe\u7247\u8f6c\u6362\u6210\u957f64\u7684\u7279\u5f81\u5411\u91cf<\/p>\n
\n>>>data = digits.images.reshape((digits.images.shape[0],-1))\n<\/code><\/pre>\n
(2)\u4f30\u8ba1\u51fd\u6570\u5bf9\u8c61<\/h3>\n
\u62df\u5408\u6570\u636e<\/strong>\uff1ascikit-learn\u5b9e\u73b0\u7684\u4e3b\u8981API\u662f\u4f30\u8ba1\u51fd\u6570\u3002\u4f30\u8ba1\u51fd\u6570\u662f\u7528\u4ee5\u4ece\u6570\u636e\u4e2d\u5b66\u4e60\u7684\u5bf9\u8c61\u3002\u5b83\u53ef\u80fd\u662f\u5206\u7c7b\u3001\u56de\u5f52\u3001\u805a\u7c7b\u7b97\u6cd5\uff0c\u6216\u8005\u63d0\u53d6\u8fc7\u6ee4\u6570\u636e\u7279\u5f81\u7684\u8f6c\u6362\u5668\u3002
\n\u4e00\u4e2a\u4f30\u8ba1\u51fd\u6570\u5e26\u6709\u4e00\u4e2afit<\/strong>\u65b9\u6cd5\uff0c\u4ee5dataset\u4f5c\u4e3a\u53c2\u6570\uff08\u4e00\u822c\u662f\u4e2a\u4e8c\u7ef4\u6570\u7ec4\uff09<\/p>\n
\n>>>estimator.fit(data)\n<\/code><\/pre>\n
\u4f30\u8ba1\u51fd\u6570\u5bf9\u8c61\u7684\u53c2\u6570<\/strong>\uff1a\u6bcf\u4e00\u4e2a\u4f30\u6d4b\u5668\u5bf9\u8c61\u5728\u5b9e\u4f8b\u5316\u6216\u8005\u4fee\u6539\u5176\u76f8\u5e94\u7684\u5c5e\u6027\uff0c\u5176\u53c2\u6570\u90fd\u4f1a\u88ab\u8bbe\u7f6e\u3002<\/p>\n
\n>>>estimator = Estimator(param1=1, param2=2)\n>>>estimator.param1\n1\n<\/code><\/pre>\n
\u4f30\u6d4b\u540e\u7684\u53c2\u6570\uff1a<\/p>\n
\n>>>estimator.estimated_param_\n<\/code><\/pre>\n
\u4e8c\u3001\u6709\u76d1\u7763\u5b66\u4e60\uff1a\u4ece\u9ad8\u7ef4\u89c2\u5bdf\u6570\u636e\u9884\u6d4b\u8f93\u51fa\u53d8\u91cf<\/h2>\n
\uff081\uff09\u8fd1\u90bb\u548c\u9ad8\u7ef4\u707e\u96be<\/h3>\n
iris\u5206\u7c7b<\/strong>\uff1a
\niris\u5206\u7c7b\u662f\u6839\u636e\u82b1\u74e3\u3001\u843c\u7247\u957f\u5ea6\u3001\u843c\u7247\u5bbd\u5ea6\u6765\u8bc6\u522b\u4e09\u79cd\u4e0d\u540c\u7c7b\u578b\u7684iris\u7684\u5206\u7c7b\u4efb\u52a1:<\/p>\n
\n>> import numpy as np\n>> from sklearn import datasets\n>> iris = datasets.load_iris()\n>> iris_X = iris.data\n>> iris_y = iris.target\n>> np.unique(iris_y)\narray([0, 1, 2])\n<\/code><\/pre>\n
\u6700\u90bb\u8fd1\u5206\u7c7b\u5668\uff1a
\n\u8fd1\u90bb\u4e5f\u8bb8\u662f\u6700\u7b80\u7684\u5206\u7c7b\u5668\uff1a\u5f97\u5230\u4e00\u4e2a\u65b0\u7684\u89c2\u6d4b\u6570\u636eX-test\uff0c\u4ece\u8bad\u7ec3\u96c6\u7684\u89c2\u6d4b\u6570\u636e\u4e2d\u5bfb\u627e\u7279\u5f81\u6700\u76f8\u8fd1\u7684\u5411\u91cf\u3002KNN(\u6700\u8fd1\u90bb)\u5206\u7c7b\u793a\u4f8b\uff1a<\/p>\n
\n#coding=utf-8   \u542b\u6709\u4e2d\u6587\u5fc5\u987b\u89c4\u5b9a\u7f16\u7801\u683c\u5f0f\nimport numpy as np\nfrom sklearn import datasets\ndef count(a1,a2,yes,no):#\u7edf\u8ba1\u6b63\u786e\u6570\u4e0e\u9519\u8bef\u6570\n    i=0\n    while i < a1.size: \n        if a1[i] != a2[i]:\n            no += 1\n        else:\n            yes += 1\n    i += 1\n    return yes,no\ndef calrate(origin,predict):\n    print "origin array:\\n %s"%(origin)\n    print "predict array:\\n %s"%(predict)\n    yes,no = 0,0\n    yes,no = count(origin,predict,yes,no)\n    rate = float(100 * yes) \/ float(yes + no) # \u6b63\u786e\u7387\n    print "\\nyes:%d no:%d \\naccuracy rate:%f%%\\n"%(yes,no,rate)\n\nif __name__ == "__main__":\n    iris = datasets.load_iris()\n    iris_X = iris.data\n    iris_Y = iris.target\n    np.random.seed(0)\n    indices = np.random.permutation(len(iris_X))\n    iris_X_train = iris_X[indices[:-10]]\n    iris_Y_train = iris_Y[indices[:-10]]\n    iris_X_test = iris_X[indices[-10:]]\n    iris_Y_test=iris_Y[indices[-10:]]\n    from sklearn.neighbors import KNeighborsClassifier\n    knn = KNeighborsClassifier()\n    knn.fit(iris_X_train,iris_Y_train)\n    predict = knn.predict(iris_X_test)\n    origin = iris_Y_test\n    calRate(origin,predict)\n<\/code><\/pre>\n
\"\"<\/a>
\n\u9ad8\u7ef4\u707e\u96be\uff08\u7ffb\u8bd1\u6709\u95ee\u9898\uff09\uff1a<\/strong>
\n\u4e3a\u4e86\u4f7f\u4f30\u8ba1\u5668\u6709\u6548\uff0c\u9700\u8981\u76f8\u90bb\u70b9\u4e4b\u95f4\u7684\u8ddd\u79bb\u5c0f\u4e8e\u67d0\u4e2a\u503cd\uff0c\u8fd9\u53d6\u51b3\u4e8e\u5177\u4f53\u95ee\u9898\u3002\u5728\u4e00\u4e2a\u7ef4\u5ea6\u4e0a\uff0c\u8fd9\u9700\u8981\u5e73\u5747n\u301c1 \/ d\u70b9\u3002\u5728\u4e0a\u8ff0k-NN\u793a\u4f8b\u7684\u4e0a\u4e0b\u6587\u4e2d\uff0c\u5982\u679c\u6570\u636e\u53ea\u6709\u4e00\u4e2a\u5177\u6709\u8303\u56f4\u4ece0\u52301\u7684\u503c\u7684\u7279\u5f81\uff0c\u5e76\u4e14\u5177\u6709n\u4e2a\u8bad\u7ec3\u89c2\u5bdf\u6837\u672c\uff0c\u5219\u65b0\u6570\u636e\u5c06\u4e0d\u4f1a\u6bd41 \/ n\u66f4\u8fdc\u3002\u56e0\u6b64\uff0c\u4e00\u65e61 \/ n\u5c0f\u4e8e\u7c7b\u95f4\u7279\u5f81\u53d8\u5316\u7684\u89c4\u6a21\uff0c\u5219\u6700\u8fd1\u90bb\u5c45\u5224\u5b9a\u89c4\u5219\u5c06\u662f\u6709\u6548\u7684\u3002
\n\u5982\u679c\u7279\u5f81\u6570\u4e3ap\uff0c\u5219\u73b0\u5728\u9700\u8981n\u301c1 \/ d ^ p\u4e2a\u70b9\u3002\u5047\u8bbe\u6211\u4eec\u5728\u4e00\u4e2a\u7ef4\u5ea6\u4e0a\u9700\u898110\u4e2a\u70b9\uff1a\u73b0\u5728\u5728P\u7ef4\u5ea6\u4e2d\u9700\u898110 ^ p\u4e2a\u70b9\u6765\u94fa\u8bbe[0\uff0c1]\u7a7a\u95f4\u3002\u968f\u7740p\u53d8\u5927\uff0c\u826f\u597d\u4f30\u8ba1\u91cf\u6240\u9700\u7684\u8bad\u7ec3\u70b9\u6570\u91cf\u5448\u6307\u6570\u589e\u957f\u3002
\n\u4f8b\u5982\uff0c\u5982\u679c\u6bcf\u4e2a\u70b9\u53ea\u662f\u5355\u4e2a\u6570\u5b57\uff088\u5b57\u8282\uff09\uff0c\u5219p-20\u7ef4\u5ea6\u4e2d\u7684\u6709\u6548k-NN\u4f30\u8ba1\u5668\u5c06\u9700\u8981\u6bd4\u6574\u4e2a\u56e0\u7279\u7f51\u7684\u5f53\u524d\u4f30\u8ba1\u5927\u5c0f\u66f4\u591a\u7684\u8bad\u7ec3\u6570\u636e\uff08\u00b11000\u57c3\u6bd4\u5b57\u8282\uff09\u3002
\n\u8fd9\u88ab\u79f0\u4e3a\u7ef4\u5ea6\u707e\u96be\uff0c\u662f\u673a\u5668\u5b66\u4e60\u7684\u6838\u5fc3\u95ee\u9898\u3002<\/p>\n
\uff082\uff09\u7ebf\u6027\u6a21\u578b\uff1a\u4ece\u56de\u5f52\u5230\u7a00\u758f\u6027<\/h3>\n
Diabets\u6570\u636e\u96c6\uff08\u7cd6\u5c3f\u75c5\u6570\u636e\u96c6\uff09
\n\u7cd6\u5c3f\u75c5\u6570\u636e\u96c6\u5305\u542b442\u4e2a\u60a3\u8005\u768410\u4e2a\u751f\u7406\u7279\u5f81\uff08\u5e74\u9f84\uff0c\u6027\u522b\u3001\u4f53\u91cd\u3001\u8840\u538b\uff09\u548c\u4e00\u5e74\u4ee5\u540e\u75be\u75c5\u7ea7\u6570\u6307\u6807\u3002<\/p>\n
\ndiabetes = datasets.load_diabetes()\ndiabetes_X_train = diabetes.data[:-20]\ndiabetes_X_test = diabetes.data[-20:]\ndiabetes_y_train = diabetes.target[:-20]\ndiabetes_y_test = diabetes.target[-20:]\n<\/code><\/pre>\n
\u6211\u4eec\u7684\u4efb\u52a1\u662f\u4ece\u751f\u7406\u7279\u5f81\u9884\u6d4b\u75be\u75c5\u7ea7\u6570
\n\u7ebf\u6027\u56de\u5f52<\/strong>
\n\u3010\u7ebf\u6027\u56de\u5f52\u3011\u7684\u6700\u7b80\u5355\u5f62\u5f0f\u7ed9\u6570\u636e\u96c6\u62df\u5408\u4e00\u4e2a\u7ebf\u6027\u6a21\u578b\uff0c\u4e3b\u8981\u662f\u901a\u8fc7\u8c03\u6574\u4e00\u7cfb\u5217\u7684\u53c2\u6570\u4ee5\u4f7f\u5f97\u6a21\u578b\u7684\u6b8b\u5dee\u5e73\u65b9\u548c\u5c3d\u91cf\u5c0f
\n\u7ebf\u6027\u6a21\u578b\uff1ay = \u03b2X+b
\nX:\u6570\u636e
\ny\uff1a\u76ee\u6807\u53d8\u91cf
\n\u03b2\uff1a\u56de\u5f52\u7cfb\u6570
\nb:\u89c2\u6d4b\u566a\u58f0\uff08bias\uff0c\u504f\u5dee\uff09
\n<\/p>\n
\nfrom sklearn import linear_model\nregr = linear_model.LinearRegression()\nregr.fit(diabetes_X_train, diabetes_y_train)\nprint(regr.coef_)\nnp.mean((regr.predict(diabetes_X_test)-diabetes_y_test)**2)#\u5747\u65b9\u8bef\u5dee\nprint(regr.score(diabetes_X_test, diabetes_y_test))#1\u4ee3\u8868\u6700\u4f73\u9884\u6d4b\uff0c0\u8868\u793ax\u4e0ey\u6ca1\u6709\u7ebf\u6027\u5173\u7cfb\n<\/code><\/pre>\n
\"\"<\/a><\/p>\n
\u6536\u7f29<\/strong>
\n\u5982\u679c\u6bcf\u4e00\u7ef4\u7684\u6570\u636e\u70b9\u5f88\u5c11\uff0c\u566a\u58f0\u5c06\u4f1a\u9020\u6210\u5f88\u5927\u7684\u504f\u5dee\u5f71\u54cd\uff1a
\n
\n\u9ad8\u7ef4\u7edf\u8ba1\u5b66\u4e60\u7684\u4e00\u4e2a\u89e3\u51b3\u65b9\u6848\u662f\u5c06\u56de\u5f52\u7cfb\u6570\u7f29\u5c0f\u52300\uff1a\u89c2\u6d4b\u6570\u636e\u4e2d\u968f\u673a\u9009\u62e9\u7684\u4e24\u4e2a\u6570\u636e\u96c6\u8fd1\u4f3c\u4e0d\u76f8\u5173\u3002\u8fd9\u88ab\u79f0\u4e3a\u5cad\u56de\u5f52\uff08Ridge Regression\uff09\uff1a
\n<\/p>\n","protected":false},"excerpt":{"rendered":"
1,231 Views1.\u524d\u8a00 \u6b63\u503c\u5bd2\u5047\u65f6\u671f\uff0c\u501f\u6b64\u673a\u4f1a\u5b66\u4e60sklearn\uff0c\u540c\u65f6\u8bb0\u5f55\u81ea\u5df1\u7684\u5b66\u4e60\u7ecf\u5386\uff0c\u4ee5\u4fbf\u5c06\u6765\u590d\u4e60 […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[26],"tags":[28,27],"class_list":["post-169","post","type-post","status-publish","format-standard","hentry","category-skl","tag-sklearn","tag-27"],"_links":{"self":[{"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/posts\/169","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=169"}],"version-history":[{"count":10,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/posts\/169\/revisions"}],"predecessor-version":[{"id":1242,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/posts\/169\/revisions\/1242"}],"wp:attachment":[{"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=169"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=169"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=169"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}